Communications of the ACM
Economic and business dimensions
The Internet Is Everywhere, but the Payoff Is Not
Credit: iStockPhoto.com
Amid the rapid diffusion of the Internet in the 1990s, many media commentators, investment analysts, and policymakers hoped the Internet would erase geographic and socioeconomic boundaries and improve the comparative economic performance in less prosperous regions.
The optimism had some merit. The economic tide was rising during this time, with everyone's fortunes rising. The best-sellers The Death of Distance1 and The World Is Flat4 were just two of the books that espoused that rosy view. The "productivity paradox"the question of whether ubiquitous use of computers would ever produce economic growthdisappeared among professional economists. Between 1995 and 2000 the economic tide was rising, as wages increased by 20% on average across the U.S.
Note the words "on average." What effect did the growth of the commercial Internet have on the economic performance of different locations in the U.S. economy? Our analysis published in the American Economic Review suggests the Internet is widespread, but its economic payoffs are not even.3
Misplaced Optimism
We conducted a multiyear study to document how businesses adopted, deployed, and improved the Internet from 1995 to 2000, a period of rapid initial investment by business. In prior research,2 we learned to look carefully at the specific type of investment. We distinguished between rapidly diffusing email and Web browsing and more advanced applications. Such advanced Internet applications were enabling productivity increases by lowering costs of communication between suppliers and customers over long distances, and these applications required skilled labor to implement and operate. Diffusion of these applications shaped productivity for many enterprises.
We were nevertheless struck by the persistence of an earlier perception: the Internet was almost everywhere but the pattern was uneven. Most large business establishments were using email and browsing by the late 1990s. However, the more we looked, the more we could see that advanced commercial uses of the Internet such as inventory management systems and online database sharing were found more often in larger cities than in smaller ones. They were also found more in data-heavy industries such as finance and wholesale distribution than they were in manufacturing, mining, and social services. At one level this makes senseadvanced Internet applications are more likely to be found in electronic parts supply and sophisticated financial operations than in landscaping firms and nursing homes. Nevertheless, there was a disparity: more prosperous cities were home to sophisticated companies that knew how to best take advantage of the new technology, and those companies (and cities) received more benefits from Internet use than other cities. There might well be payoffs from the Internet, but they were not evenly shared.
The Payoff Puzzle
Our more recent study combined information on business Internet usage with U.S. national-level wage data. We studied from the period 1995 to 2000 using several large datasets, including the Quarterly Census of Employment and Wages that gives county-level information on average weekly wages and employment, and the Harte Hanks Market Intelligence Computer Intelligence Technology Database of survey information about how firms use the Internet. In total, we included relevant data from over 85,000 private establishments with more than 100 employees each.
We found that business adoption of Internet technologies correlated with wage and employment growth in only 163 of the over 3,000 counties in the U.S. All of these counties had populations above 150,000 and were in the top quarter of income and educational achievement before 1995. Between 1995 and 2000, they showed a 28% average increase in wages, compared with a 20% increase in other counties, where the Internet appears to not have had much impact on economic performance (see the accompanying figure).
In short, the Internet appears to have exacerbated regional wage inequality, explaining over half the difference in wage growth between the 6% of counties that were already well-off and all other counties.
Why did the Internet make such big waves in these few areas? Three explanations are possible:
- Big cities have needed communications infrastructure and are home to firms capable of making capital investments that allow the Internet to enhance what such sophisticated firms are already doing.
- A phenomenon called "skill-biased technical change" might be under way, in which new technologies raise the wages only of skilled workers, while unskilled workers cannot use the new technology and do not receive the necessary training to benefit from its deployment through wage gains.
- Cities have denser labor markets, and better communication between supply and demand. Programmers with skills in the language of main-frames, such as COBOL, and the language of the Web, such as HMTL and Perl lived in (or moved to) big cities, where more firms would be looking for such workers. Similarly, a range of other inputs that would make Internet adoption more valuable is more available in cities.
The Internet may allow firms in rural Iowa to reach new customers on Wall Street but it also allows Wall Street banks to reach investors in rural Iowa.
Each explanation is plausible, and probably explains a piece of the story. However, existing data does not allow us to distinguish between the explanations, and so the payoff puzzle remains.
Policy
The Internet has been widely diffused and used: in this, our results echo the popular optimism. Also, the benefits are not limited to the locations that dominate supply of equipment and software, such as Boston, Seattle, New York, and Silicon Valley.
However, in contradiction to popular optimism, our data does not show any evidence of improvement in the comparative economic performance of isolated locations or less dense locations. The Internet may allow firms in rural Iowa to reach new customers on Wall Street but it also allows Wall Street banks to reach investors in rural Iowa. Our study shows that benefits from Internet adoption such as increased wages are, on balance, more likely to show up in New York City than in rural Iowa. The predicted "leveling" effects of the Internet have not materialized, and it is doubtful that short-run investment in Internet infrastructure (such as extending access to broadband) will generate immediate economic benefits for areas that have otherwise low capital investment, inadequate labor skills, and market conditions.
Conclusion
These results have important implications for public policy. Many lawmakers have argued for government to subsidize the expansion of the Internet into poor and isolated regions. For example, the 2009 economic stimulus package allocated over $7 billion for broadband expansion in under-served areas, under the assumption that creating such infrastructure will raise wages and income. Our results suggest such Internet infrastructure investments by themselves are unlikely to raise wages in poor and isolated regions, at least in the short run. Such investments might be justified to strengthen education, increase civic engagement, or promote health and safety. And, over 20 to 30 years expanding Internet access might lead to more widespread economic gains. However, there is little evidence to suggest a short-term payoff in the form of increased local wages.
References
1. Cairncross, F. The Death of Distance. Harvard University Press, Cambridge, MA, 1997.
2. Forman, C., Goldfarb, A., and Greenstein, S. How did location affect adoption of the commercial Internet? Global village vs. urban leadership. Journal of Urban Economics 58, 3 (Mar. 2005), 389420.
3. Forman, C., Goldfarb, A., and Greenstein, S. The Internet and local wages: A puzzle. American Economic Review 102, 1 (Jan. 2012), 556575.
4. Friedman, T.L. The World is Flat: A Brief History of the Twenty-First Century. Farrar, Straus, and Giroux, New York, 2005.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2012 ACM, Inc.
No entries found